*Article* **Enhanced Stability and Mechanical Properties of a Graphene–Protein Nanocomposite Film by a Facile Non-Covalent Self-Assembly Approach**

**Chunbao Du <sup>1</sup> , Ting Du <sup>1</sup> , Joey Tianyi Zhou <sup>2</sup> , Yanan Zhu <sup>1</sup> , Xingang Jia <sup>1</sup> and Yuan Cheng 3,4,\***


**Abstract:** Graphene-based nanocomposite films (NCFs) are in high demand due to their superior photoelectric and thermal properties, but their stability and mechanical properties form a bottleneck. Herein, a facile approach was used to prepare nacre-mimetic NCFs through the non-covalent selfassembly of graphene oxide (GO) and biocompatible proteins. Various characterization techniques were employed to characterize the as-prepared NCFs and to track the interactions between GO and proteins. The conformational changes of various proteins induced by GO determined the film-forming ability of NCFs, and the binding of bull serum albumin (BSA)/hemoglobin (HB) on GO's surface was beneficial for improving the stability of as-prepared NCFs. Compared with the GO film without any additive, the indentation hardness and equivalent elastic modulus could be improved by 50.0% and 68.6% for GO–BSA NCF; and 100% and 87.5% for GO–HB NCF. Our strategy should be facile and effective for fabricating well-designed bio-nanocomposites for universal functional applications.

**Keywords:** graphene; nanocomposite film; film-forming ability; stability; mechanical properties

#### **1. Introduction**

Two-dimensional (2D) nanomaterials have recently opened a new era for flexible devices owing to their exotic electronic and optical properties [1–3]. Graphene is an emerging constituent for 2D nanomaterials, and graphene films hold great potential for meeting various intellectualized functionalities [4–6]. However, the mechanical properties of pure graphene films have significant flaws, such as limited flexibility and stability [7,8]. Reinforcing components are usually added to produce nanocomposite films (NCFs) to improve overall characteristics, which opens new avenues for graphene's use. Synthetic polymers are used in most graphene-based NCFs owing to their superior designability and usefulness [9,10]. However, synthetic polymers do not easily decompose naturally, resulting in considerable solid waste [11,12]. Therefore, the present trend is to develop environmentally friendly graphene-based NCFs to reduce carbon emissions and allow more recycling of materials.

Biomacromolecules (BMMs), which are indispensable for in vivo life, including proteins, polypeptides, enzymes, DNA, RNA, lipids, and polysaccharides, are being used in in vitro applications because of their exceptional functionality and biodegradability [13–17]. When included in NCFs with nanomaterials that have the desired photoelectric and thermal properties, multifarious applications are available, such as biosensors, artificial tissue, information storage, and drug delivery [18–20]. Various studies report that integrating

**Citation:** Du, C.; Du, T.; Zhou, J.T.; Zhu, Y.; Jia, X.; Cheng, Y. Enhanced Stability and Mechanical Properties of a Graphene–Protein Nanocomposite Film by a Facile Non-Covalent Self-Assembly Approach. *Nanomaterials* **2022**, *12*, 1181. https://doi.org/10.3390/ nano12071181

Academic Editor: Orietta Monticelli

Received: 4 March 2022 Accepted: 31 March 2022 Published: 1 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

BMMs and graphene has already been done, forming some novel composites will multiple applications [20–25]. For example, Liu et al. employed a simple method to prepare low-cost graphene and silk-based pressure sensor, which could be used as artificial skin to monitor the pressure of the human body in real-time [26]. In another case, Chu et al. fabricated a hybrid scaffold using graphene oxide (GO) and an acellular dermal matrix, promoting cell proliferation in the wounds of diabetics [27]. Recently, Chang et al. loaded a heat shock protein 90 inhibitor NVP-AUY922 on a GO-based GO/BaHoF5/PEG nanocomposite to perform sensitized photothermal therapy (PTT). The achieved nano-platform, GO/BaHoF5/PEG/NVP-AUY922, had excellent biocompatibility and made tumor cells more sensitive to hyperthermia, which could promote the development of low-laser-hazard PTT [28]. In the work of Zhao et al., photomodule single-layer reduced graphene oxide (rGO) has been organized into a well-defined multilayer stack with the help of amyloidlike protein aggregates [29]. The as-fabricated hybrid film reliably adheres to the plastic substrate with robust interfacial adhesion. The sensitive photothermal effect of rGO in the bilayer film can be initiated with a blue laser from 100 m away, indicating that the combination of GO with BMMs exhibited great potential in remote light control of robots and devices. Our previous work has summarized the bio–nanomaterial interaction mechanisms at the molecular level of some typical 2D nanomaterials and BMMs, including non-covalent and covalent interactions, and proposed the challenges for the future development of 2D materials and biomacromolecules [30]. Despite significant advances, insufficient attention has been devoted to the stability of graphene-based NCFs in applicable environments involving acidity, alkalinity, salt, heat, and so on. Furthermore, the production processes of NCFs with graphene and biomacromolecules at the molecular level, including the species, conformations, film-forming ability, and mechanical characteristics, need to be investigated further. As a result, there are still major opportunities in, and obstacles to, extending more general biomacromolecules, including understanding and controlling molecular pathways.

In this work, after considering the desirability of simplicity, low cost, and reproducibility for BMMs to be used in scaled-up applications, ordinary and commercialized proteins, i.e., bovine serum albumin (BSA) and hemoglobin from bovine-blood (HB), were chosen for assembly with GO to fabricate NCFs (i.e., GO–BSA NCF and GO–HB NCF; Figure 1a). Using lysozyme (Lyz)-formed NCF (i.e., GO–Lyz NCF) for a comparison, the film-forming abilities of the NCFs were investigated comprehensively by tracking the experimental progress and analyzing the microstructures. X-ray photoelectron spectrometry (XPS), scanning electron microscopy (SEM), and circular dichroism (CD) spectroscopy, along with theoretical simulations, were adopted to reveal the binding mechanism. Moreover, the stability, thermostability, and mechanical properties were investigated by dissolution experiments, differential scanning calorimetry (DSC), and nanometer indentation. Although the structures and properties of BSA, HB, and Lyz are different, consistent stability and improved mechanical strength were achieved, which might provide some inspiration for fabricating other stable nanocomposites.

**Figure 1.** Scheme of self-assembly of GO with BSA/HB for GO–BSA and GO–HB NCF (**a**). Possible binding mechanism of a GO sheet with BSA/HB and the fabrication of a NCF (**b**).

#### **2. Materials and Methods**

#### *2.1. Experimental Materials*

GO with a thickness of 1.2–5 nm and a lateral size of 50 nm to 3 µm was purchased from Hangzhou Nano-Mall Technology Co., Ltd., Hangzhou, China. BSA, HB, and Lyz were supplied by Macklin. NaCl (AR, >99.5%), HCl (AR, 36.0–38.0%), and NaOH (AR, >99.6%) were products of Sinopharm Chemical Reagent Co., Ltd., Shanghai, China.

#### *2.2. Fabrication of NCF*

The fabrication of NCF was conducted by using simple vacuum filtration. Typically, protein (i.e., BSA, HB, or Lyz) was dissolved in the deionized water to form the homogeneous solution with 4 mg mL−<sup>1</sup> . After that, the aqueous GO solution (4 mg mL−<sup>1</sup> , 5 mL) was added to the protein solution (5 mL) after the ultrasonic treatment of the aqueous GO solution at 100 W, at 20 ◦C, for 30 min. Afterward, the mixed solution was stirred at room temperature (26.4 ◦C) for 16 h to complete the self-assembly process. Then, the mixed solution was poured into the filter flask to remove the water and obtain the wettish NCF. After drying at 30 ◦C in a vacuum drying oven for 2 h, the dried GO–BSA, GO–HB, and GO–Lyz NCF were kept in a desiccator. The preparation method of the GO film was the same as that of the above NCFs without adding protein. Notably, the structure and characteristics of the composite films varied depending on the proportions of GO and protein [31]. During the experiments, other ratios of GO and protein were also tested, but the films obtained were all poor. When the amount of GO was higher than protein, the film was brittle. In the opposite case, the film was thin and difficult to remove from the substrate.

#### *2.3. Analysis and Characterization*

SEM (FEI Talos F200X) was used to observe the morphology and structure of each film at a high voltage of 10.0 kV. Chemical surface characterization was performed by XPS (Shimadzu Kratos, Manchester, United Kingdom) with monochromatic Al Ka radiation (1486.6 eV); the deconvolution method using Gaussian and Lorentz curve fittings was employed to conduct the semiquantitative analysis of the elements. The thermal properties of films were characterized by DSC (Nestal DSC214, Selb, Germany). The CD spectroscopy experiments were carried out using a CD spectrometer (Applied Photophysics Ltd. Chiras-

can, Leatherhead, UK); the GO–BMM compound solution after self-assembly was diluted by a factor of 27; CD spectra were obtained by scanning the diluted GO–BMM compound solution and deducting the background of the GO solution. Secondary structures of BSA and HB were determined by fitting the far-UV CD data using CDNN algorithms. An optical microscope (Olympus BX51, Tokyo, Japan) was used to observe the morphology of GO and GO–BMM solutions at room temperature. The mechanical properties of NCFs were characterized by a nanometer indentation instrument (UNHT) at room temperature (28.3 ◦C) with five random positions (RP) and the compression rate of 1 mm min−<sup>1</sup> ; *F*<sup>m</sup> (mN) and *h* (nm) were confirmed in the obtained curved; *H*IT (GPa) and *E*\* (GPa) were calculated by the supporting analysis software. The stability of films was measured in separate NaCl, HCl, and NaOH aqueous solutions, each with a concentration of 0.1 M.

#### **3. Results and Discussion**

#### *3.1. Structural Property of NCFs*

The fabrication of NCFs was conducted using simple vacuum filtration (Figure 1a). Protein (i.e., BSA, HB, and Lyz; see Figure S1a–c) was dissolved in the deionized water to form the homogeneous solution and then mixed with the ultrasonically treated GO aqueous solution. After stirring, the mixed solution was poured into a filter flask to remove the water and obtain the wettish film. Finally, the GO–BSA, GO–HB, and GO–Lyz NCFs were obtained after drying. The most spread out BSA, HB, and Lyz were in the three dimensions of the crystalline state was about 14 <sup>×</sup> <sup>8</sup> <sup>×</sup> 5 nm<sup>3</sup> , which was far less than the size of GO sheet (50 nm–3 µm) [32]. BSA, HB, and Lyz were more likely to be bound on the surface of a GO sheet rather than the edge. The interactions between the GO sheet and BSA/HB/Lyz were dominated by the multiple non-covalent interactions (Figure 1b). Notably, because protein molecules could not be trapped by the filter membrane, the pure protein film could only be obtained by solvent evaporation and could not be obtained using vacuum filtration. GO is a 2D material with a high aspect ratio that is useful for adsorbing protein molecules as a skeleton when creating films, and its outstanding mechanical strength is very advantageous [33,34]. To unravel the protein–GO interaction mechanism, specific experiments were designed, and the results obtained are discussed step by step as follows.

The sizes (layers, transverse, and longitudinal direction) and properties (functional groups and groups density) of GO were detected first. The GO used in this work was prepared through a typical Hummers sonication method [35]. The thickness and lateral size of the GO sheet were 1.2–5 nm and 50 nm to 3 µm, respectively. Therefore, this ensured that the GO sheet was stretching rather than crimping into spherical particles [36]. XPS data (Figure S1d) showed the main elements of GO were C and O, and there was very little N. The functional groups of the GO sheet included the following percentages of the total carbon: carbonyl (C=O), 2.0%; hydroxyl (-OH), 46.5%; carboxyl (-C=(O)-OH), 2.5% (Figures 2a and S1e). Those percentages imply that the graphene structure in GO sheet largely remained intact. The typical elements in BSA, HB, and Lyz other than C and O were N and S. Figure 2b,c shows the N 1s and S 2p XPS survey spectra of GO–BSA, GO–HB, and GO–Lyz NCF, further indicating the existence of BSA, HB, and Lyz. Moreover, the different atomic concentrations of each element in these films could also support the formation of these NCFs (Figure 2d). The polar binding sites of GO sheets usually existed on the edges and in the defects on the surface. Therefore, BSA/HB/Lyz was more likely to be bound on the surface of the GO sheet rather than the edge due to the huge difference between the size of any of these proteins and that of GO sheets.

The insets in Figures S2a and 3a–c show the megashapes of GO film, GO–BSA NCF, GO–HB NCF, and GO–Lyz NCF. GO–BSA and GO–HB NCFs had complete structures and exhibited bendability, whereas the GO film and GO–Lyz NCV were easily broken quickly. The reasons for this phenomenon were associated with the properties of these proteins and their binding situations with GO sheet. To further investigate the binding situations between the GO sheet and these proteins, SEM was employed to discover the surface microtopography. For the GO film, due to the polar sites on the surface and corner

of the GO sheet, it was not easy to accomplish the *π*–*π* tight stacking in the GO sheets. That caused the GO film to have an uneven surface with great roughness (Figure S2a). On the contrary, GO sheet–BSA and GO sheet–HB compounds were achieved by adequate self-assembly, and their surfaces showed relatively good uniformity (Figure 3a,b). Thereinto, BSA and HB played the role of plasticizer to adjust the interfacial compatibility of GO. However, for the GO–Lyz NCF, its microscopic surface was the same as that of the GO film (Figure 3c), which meant the effect of this approach was negligible. These differences were also reflected in the GO sheet–protein compound solutions macroscopically. For both GO sheet–BSA and GO sheet–HB, the compound solutions, after a 16 h self-assembly process before suction filtration, were stable suspensions without precipitation or aggregation (Figure S3b,c). Most interestingly, both GO sheet–BSA and GO sheet–HB compound solutions were stable in polar aqueous solutions, which were more stable than the GO solution (Figure S3a), indicating that the external surfaces of GO sheet–BSA and GO sheet–HB compounds are also polar. Nevertheless, the external surface of GO sheet–Lyz compound was hydrophobic, and aggregation of the GO sheet–Lyz compound occurred through the spontaneous hydrophobic interactions, causing the apparent precipitation phenomenon of GO sheet–Lyz in the aqueous phase (Figure S3d). The stabilities of these compounds in aqueous solutions were closely related to the properties of the obtained NCFs.

SEM images of the internal cross-section were more valid evidence to confirm the above analysis. As shown in Figure 3d,e, both GO–BSA NCF and GO–HB NCF displayed a prominent layered hierarchical structure of natural nacre. Moreover, their compactness and smoothness in section micromorphology were better than those of GO and GO–Lyz NCFs when compared with the inter-layer gaps of the latter films (Figures 3f and S2b). Although GO sheets were stable in the aqueous phase due to their polar sites, there were inevitably irregular gaps between GO sheets among the *π*–*π* tight stacking. By contrast, BSA and HB had better adhesion to GO sheets to facilitate the mutual attraction and fill the gaps for the dense structures. For Lyz, the structural change induced by the GO sheet was not beneficial for the tight and homogeneous binding of GO sheet–Lyz, and the instability of GO sheet– Lyz in an aqueous solution also caused inhomogeneity of GO–Lyz NCF. The excellent interfacial compatibility of GO sheet–protein compounds contributed to enhancing the mechanical properties of NCFs. Thereinto, the conformational changes of proteins induced by GO sheet were essential. Considering film-forming ability, GO–BSA NCF and GO–HB NCF are better candidates than GO–Lyz NCF for practical applications. Furthermore, using the SEM images of interior cross-sections, the thicknesses of GO–BSA and GO–HB NCF were determined to be around 2.1 and 2.2 µm, respectively.

The mechanical properties of GO–BSA and GO–HB NCF could also be reflected by the conformational changes of BSA and HB. To prove the secondary structure changes in BSA/HB induced by GO sheets, CD spectra of BSA and HB aqueous solutions with and without the addition of GO were obtained. The secondary structures (i.e., *α*-helix, *β*-pleated sheet, *β*-turn, and random coil) of proteins were confirmed by the positions of *α*-helixes (222 and 208 nm positive peaks, 192 nm negative peak), *β*-pleated sheets (217–218 nm positive peaks, 195–198 nm negative peak), *β*-turns (220–230 nm weak positive peaks, 180–190 nm strong positive peaks, 205 nm negative peak), and random coils (198 nm positive peaks, 220 nm negative peak) [37]. Figure 2e showed that there were apparent changes in the secondary structures of BSA and HB before and after binding of GO sheets, indicating that GO had the apparent effect on the structures of BSA and HB. After analyzing the data of contents of the secondary structures (Table 1), the changes in the secondary structures in BSA and HB showed the same pattern, which was a decrease in *α*-helixes and increases in *β*-pleated sheets (antiparallel and parallel), *β*-turns, and random coils. Before introducing GO sheets, *α*-helixes predominated in BSA and HB with the contents of 50.7 and 46.0%, respectively. After interactions with GO sheets, the *α*-helix contents of BSA and HB decreased to 14.6 and 16.4%, respectively. Our previous work has revealed the binding mechanism of the *α*-helix fragments of BSA with graphene by using molecular dynamics simulations [38]. The adsorption of an *α*-helix on the surface of graphene induces

a transition from to the 310-helix structure, which was reflected in the substantial increase in random coils from 24.1 to 42.2% for BSA in this work. This induction mode might also work for HB because the content of the random coils of HB increased from 26.5 to 41.4%. For the *β*-pleated sheets (antiparallel and parallel), the tiled state on GO sheets was more stable due to the interactions of more binding sites, which was in accordance with the molecular dynamics simulations of our previous work that showed graphene was advantageous to the stability of *β*-pleated sheets [39]. The increase in *β*-turns always accompanied the increase in *β*-pleated sheets. Therefore, the increases in contents of the *β*-pleated sheets (antiparallel and parallel) and random coils in BSA and HB were beneficial for the self-assembly of GO sheet with BSA and HB.

**Figure 2.** C 1s high magnification of films (**a**). N 1s high magnification of GO film (**b**). S 2p high magnification of films (**c**). The atomic compositions of films (**d**). CD spectra of BSA and HB with and without GO induction (**e**). DSC curves of GO film, BSA, HB, GO–BSA NCF, and GO–HB NCF with a heating rate of 5 ◦C min−<sup>1</sup> in N<sup>2</sup> flow from 25 to 100 ◦C (**f**).

**Figure 3.** SEM images of surface (**a**–**c**); photographs with a uniform diameter of about 5 cm (insets of (**a**,**c**,**e**)). Internal cross-sections (**d**–**f**) of GO–BSA NCF, GO–HB NCF, and GO–Lyz NCF.


**Table 1.** The contents of the secondary structure elements of BSA and HB with and without GO.

#### *3.2. Stability of NCFs*

The thermostability of films is very important to determine their applications at different temperatures. The thermostability of GO film, GO–BSA NCF, and GO–HB NCF was characterized by DSC. As shown in Figure 2f, heat release of the GO film proceeded the increase in temperature, indicating that the GO sheet was very sensitive to heat. After binding with BSA and HB, only transient heat release occurred in GO–BSA and GO–HB NCFs, and then the heat flows were maintained within a stable range, implying that GO–BSA and GO–HB NCF were thermostable. It is not difficult to see that BSA and HB also exhibited continuous heat release and absorption around 60–70 ◦C due to their conformational changes with the temperature change. The glass transition is the transition of amorphous material from a glassy state to a high elastic state. As shown in the inset of Figure 2f, the glass-transition temperatures of GO–BSA and GO–HB NCFs were confirmed to be 25.3 and 25.4 ◦C, respectively, which belong to the scope of room temperature. This indicates that the GO–BSA and GO–HB NCF could be kept in high elastic states at room temperature and remain stable. It was advantageous for GO–BSA and GO–HB NCFs to fully utilize their flexibility for stability. Before the self-assembly process of GO with BSA or HB, the conformational changes of BSA or HB were completed, so the film-forming process would not induce a conformational change in BSA or HB. That is to say, the filmforming process would only involve the self-assembly of "GO sheet-BSA/HB compounds." Combined with the results of CD spectra, it could be concluded that the introduction of BSA or HB on the surface of a GO sheet was helpful to improving the thermostability, which is attributed to the increased contents of the *β*-pleated sheets (antiparallel and parallel) and random coils.

The stability of films in various complex environments is also essential for their actual applications. It has been confirmed that the functional groups of the GO sheet were OH and –COOH. Even so, there were many hydrophobic areas on the surface of the GO sheet. The formation of the GO film was caused by the polar and non-polar interactions of GO sheets. The polar interactions included hydrogen bonding of –OH with –OH, –OH with –COOH, and –COOH with –COOH; non-polar interactions were *π*–*π* stacking. The GO film was barely stable in the aqueous phase and was dissolved partly after 7 days (Figure S4a). After ultrasonication, the mutual attraction of GO sheets could not conquer the destructive effect from outside that dissolved GO film easily. –COOH↔–COO−+H<sup>+</sup> was a dynamic equilibrium process and water could break the hydrogen bonding to a certain degree. GO films were stable in acidic, alkaline, and saline environments for a standing time of 7 days (Figure S4b–d). However, they were unstable under ultrasonication in alkaline and saline environments because the films were dissolved easily, which we attribute to different dissolution mechanisms. The increase in –COO− groups facilitated the electrostatic repulsion in an alkaline environment, and the saline ions destroyed the electrostatic attraction in a saline environment. Both cases were not beneficial for the stability the GO films. By contrast, GO films were still stable in an acidic environment, even under ultrasonication, because the increase in –COOH groups could tremendously enhance the mutual attraction of GO sheets. After determining the stability and instability mechanisms of GO films in the above environments, the assembly mechanisms of GO–BSA and GO–HB NCFs were also analyzed. The main functional groups of BSA and HB are

–NH3, –COOH, hydrophobic chains (benzene ring and alkyl chain), and other polar groups (–OH, –C(=O)–NH–). The formation of GO–BSA and GO–HB NCFs meant self-assembly of GO sheet–BSA and GO sheet–HB compounds, respectively, which still involved multiple non-covalent interactions, as in the formation of GO films. The process was markedly different for BSA and HB, in types of interactions and binding strengths, owing to their uniqueness. Therefore, the stability of GO–BSA and GO–HB NCFs differed significantly in different environments. GO–BSA NCF was very stable in aqueous, acidic, alkaline, and saline environments with or without ultrasonic treatment, indicating that the favorable interactions for film-forming were far stronger than the adverse interactions (Figure 4a–d). There was no denying that the confirmation of interactions is very complex and needs precision instruments and testing [40,41]. GO–HB NCF was very stable in acidic and saline environments regardless of ultrasonic treatment (Figure 4f,h), but it can dissolve easily in aqueous and alkaline environments with ultrasonic treatment (Figure 4e,g). The stability of these films will determine their ranges of application, which is the case for all BMMs with a unique characteristics.

**Figure 4.** The stability of the GO–BSA NCF in aqueous (**a**), HCl (**b**), NaOH (**c**), and NaCl (**d**) solutions at room temperature. Stability of GO–HB NCF in aqueous (**e**), HCl (**f**), NaOH (**g**) and NaCl solution (**h**) with different treatments at room temperature.

#### *3.3. Mechanical Properties of NCFs*

The mechanical properties of GO film, GO–BSA NCF and GO–HB NCF were characterized by nanoindentation with five random positions (RP) under the maximum applied load (*F*m, mN) of 5 mN. Generally, there are three stages, namely, loading, holding, and unloading (Figure 5a), which are very apparent in *H*IT-*h* curves. There were significant differences in the tracks of five *H*IT-*h* curves of the GO films, indicating that the uniformity of GO films was relatively poor (Figure 5b), which is in accordance with the results of SEM in Figure S2a: the surfaces of GO films were uneven with large roughness. However, the tracks for GO–BSA and GO–HB NCFs (Figure 5c,d) were nearly identical, indicating that the uniformity of GO–BSA NCF and GO–HB NCF is much better than that of GO film. In addition, the corresponding indentation hardness (*H*IT, GPa) and equivalent elastic modulus (*E*\*, GPa) of each *H*IT-*h* curve can be calculated. Figure 5e,f shows the *H*IT and *E*\* of each RP of a GO film, GO–BSA NCF, and GO–HB NCF, and there are no abrupt values, indicating these films had good structural homogeneity and no flaws. Numerous studies have strived to improve the mechanical properties of GO-based films because their limited mechanical properties have hindered practical applications [42–44]. Here, the average *H*IT and *E*\* of GO–BSA NCF were 0.12 and 2.7 GPa, respectively; the average *H*IT and *E*\* of GO–HB NCF were 0.16 and 3.0 GPa, respectively. Compared with average the *H*IT (0.08 GPa) and *E*\* (1.6 GPa) of GO film, the mechanical properties of GO–BSA and GO–HB NCF were very much improved. Under *F*<sup>m</sup> of 5 mN, the depths of indentation of GO–BSA and GO–HB NCF were around 1300 and 1200 nm, which were lower than that of GO film (1600–2000 nm), indicating that the existence of BSA and HB in interlayers of films

could store the stress. BSA and HB served as filling agents and plasticizers to complement the stretchability of GO. In the work of Li et al., the functional groups on GO sheets had a significant influence on the mechanical properties of a GO–silk-based nanocomposite, and the oxygen-containing groups of GO could form hydrogen bonding with silk fibroins at the interface to improve the adhesive force [45]. A similar principle applies for BSA and HB, because BSA or HB could bind to the surface of a GO sheet strongly through hydrogen bonding, in addition the spontaneous hydrophobic interactions, which would shield the sheet from water molecules surrounding GO–BSA NFC or GO–HB NCF, thereby stabilizing their structures [45]. Therefore, the resistance to instantaneous and continuous external forces in GO–BSA and GO–HB NCF is significantly improved over that of GO films, meaning that GO–BSA and GO–HB NCFs exhibit enhanced applicability. In the previous work of Shao and Fan et al., bacterial cellulose and chitosan were fabricated with GO to improve the mechanical properties of GO–bacterial cellulose and GO–chitosan NCF, respectively, and the assembly process only involved hydrogen bonding and electrostatic interactions [46,47]. Compared with Shao's work with an *E*\* of 0.5 GPa, GO–BSA NCF and GO–HB NCF had significant advantages in terms of mechanical properties [46]. In addition, although the *H*IT (0.40 GPa) and *E*\* (6.5 GPa) of GO–chitosan NCF in Fan's work were much higher than those of GO–BSA and GO–HB NCF, GO–chitosan NCF was extremely unstable in an acidic environment, which will handicap its applicability [47]. Therefore, combining GO and well matched BMMs is an effective and fantastic strategy to construct nacre-like NCFs with good stability and enhanced mechanical properties.

**Figure 5.** Typical scheme of nanoindentation load–displacement curve (**a**). Representative nanoindentation load–displacement curves for GO film (**b**), GO–BSA NCF (**c**), GO–HB NCF (**d**) with five RP. The corresponding *H*IT (**e**) and *E*\* (**f**) of the GO film, GO–BSA NCF, and GO–HB NCF.

#### **4. Conclusions**

In summary, GO–BMM-based NCFs with nacre-mimetic structures were fabricated with GO and proteins through a green and straightforward non-covalent self-assembly process. The sequences and conformational features of BSA, HB, and Lyz determined their film-forming ability with GO sheets, implying that GO–BSA and GO–HB, which are stable compounds in an aqueous solution, are outstanding candidates for fabricating stable NCFs. The GO sheet induced increases in the presence of *β*-pleated sheets, *β*-turns, and random coils in BSA and HB, along with a decrease in the presence *α*-helixes, which was more beneficial for the NCFs with dense and uniform microstructures. Compared with the GO film, GO–BSA and GO–HB NCF exhibited good thermostability below 100 ◦C, and remained remarkably stable in acidic and saline environments. GO–BSA NCF could be kept stable in an alkaline environment, which endows it with broader application potential. Moreover, GO–BSA and GO–HB NCF exhibited significant advantages through appreciable

HIT enhancements of 50.0% and 100%; and enhancements in *E*\* by 68.6% and 87.5%, respectively. The binding of BMMs into interlayers of 2D nanomaterials could synergistically provide enhancements while maintaining the films' respective characteristics, making them promising for flexible devices.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/nano12071181/s1. Figure S1: The crystal structures of BSA (a, PDB code 4F5S), HB (b, PDB code 1FN3), and Lyz (c, PDB code 253L) with the style of newcartoon; wide scan of XPS survey spectra of films (d); atomic concentration of C in GO film (e). Figure S2: SEM images of surfaces (a); photographs with uniform diameter of about 5 cm (insets of a). Internal cross section (b) of GO film. Figure S3: Optical microscope photographs of the GO solution (a), GO–BSA compound solution (b), GO–HB compound solution (c) and GO–Lyz compound solution (d); Figure S4: The stability of GO film in aqueous (a), HCl (b), NaOH (c), and NaCl (d) solutions at room temperature.

**Author Contributions:** Conceptualization, C.D. and Y.C.; methodology, C.D.; formal analysis, C.D. and Y.C.; investigation, C.D. and T.D.; resources, C.D. and Y.C.; writing—original draft preparation, C.D. and Y.C.; writing—review and editing, J.T.Z., Y.Z. and X.J.; project administration, C.D. and Y.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Natural Science Foundation of China (22002117), the Natural Science Foundation of Shaanxi Province, China (2021JQ-585), and the Scientific Research Program of Shaanxi Provincial Education Department (20JK0839). J.T.Z. appreciates the financial support from the RIE2020 AME Programmatic Grant A18A1b0045 funded by A\*STAR-SERC, Singapore.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** J.T.Z. is grateful for the support from the Agency for Science, Technology, and Research (A\*STAR), A\*STAR Computational Resource Centre, Singapore (ACRC); and the National Supercomputing Centre, Singapore (NSCC). C.D. thanks the Modern Analysis and Testing Center of Xi'an Shiyou University and Shiyanjia Lab (www.shiyanjia.com (accessed on 1 November 2021)) for characterizations.

**Conflicts of Interest:** Authors declare no conflict of interest.

#### **References**


**Suhui Wang, Xu Zhang, Yi Wang , Tengxiao Guo \* and Shuya Cao \***

State Key Laboratory of NBC Protection for Civilian, Beijing 102205, China; wangsuhui1995@163.com (S.W.); m4a1hitman@sina.com (X.Z.); wangyi102205@sina.com (Y.W.)

**\*** Correspondence: guotengxiao@sklnbcpc.cn (T.G.); caoshuya@sklnbcpc.cn (S.C.)

**Abstract:** As the core device of the miniature quantum dot (QD) spectrometer, the morphology control of the filter film array cannot be ignored. We eliminated strong interference from additives on the spectrum of a long-wave infrared (LWIR) QD filter film by selecting volatile additives. This work is significant for detecting targets by spectroscopic methods. In this work, a filter film with characteristic spectral bands located in the LWIR was obtained by the natural evaporation of QD ink, which was prepared by mixing various volatile organic solvents with HgSe QD–toluene solution. The factors affecting the morphology of HgSe LWIR films, including ink surface tension, particle size, and solute volume fraction, were the main focus of the analysis. The experimental results suggested that the film slipped in the evaporation process, and the multilayer annular deposition formed when the surface tension of the ink was no more than 24.86 mN/m. The "coffee ring" and the multilayer annular deposition essentially disappeared when the solute particles were larger than 188.11 nm. QDs in the film were accumulated, and a "gully" morphology appeared when the solute volume fraction was greater than 0.1. In addition, both the increase rate of the film height and the decrease rate of the transmission slowed down. The relationship between film height and transmission was obtained by fitting, and the curve conformed to the Lambert–Beer law. Therefore, a uniform and flat film without "coffee rings" can be prepared by adjusting the surface tension, particle size, and volume fraction. This method could provide an empirical method for the preparation of LWIR QD filter film arrays.

**Keywords:** HgSe QD; long-wave infrared; evaporated film; morphology

#### **1. Introduction**

Nanomaterial inkjet printing technology is a cutting-edge technology for the microdistribution and precise printing of ink droplets by controlling the nozzle voltage, air pressure, platform temperature, and motion trajectory, which can achieve the high-precision patterned deposition of nanomaterials. This technology has attracted extensive attention in the fields of display panel printing [1–4], microelectronic component fabrication [5–7], and flexible printing [8–10] in recent years due to the advantages of rapidity, convenience, and low cost.

As is known to all, semiconductor QDs are synthetic nanomaterials. When the size of a semiconductor quantum dot is smaller than or comparable to the exciton Bohr radius in all three directions, the electron motion is confined and forms a split energy level [11–21]. Therefore, it has many unique optoelectronic properties different from those of bulk materials, such as broad absorption spectra, narrow symmetrical emission spectra, and large Stokes shifts. Its spectrum can be tuned by adjusting various parameters, such as the synthesis time, material ratio, and core–shell structure [22,23]. In addition to the QD's unique optical properties, it also has the advantages of low cost, multiple types, and easy integration. Therefore, people have used QDs as filter materials to prepare visible-light

**Citation:** Wang, S.; Zhang, X.; Wang, Y.; Guo, T.; Cao, S. Influence of Ink Properties on the Morphology of Long-Wave Infrared HgSe Quantum Dot Films. *Nanomaterials* **2022**, *12*, 2180. https://doi.org/10.3390/ nano12132180

Academic Editor: Orion Ciftja

Received: 1 June 2022 Accepted: 21 June 2022 Published: 24 June 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

(Vis) and near-infrared (NIR) filter film arrays and QD micro-spectrometers [24–29]. The existing filter films and their working bands are shown in Table 1.

**Table 1.** Existing filter films and their working bands.


Using a QD filter film array prepared by inkjet printing technology as the spectroscopic element of a micro-spectrometer is an effective method to miniaturize the spectrometer. The uniformity of the filter film directly determines the error when the detector detects the light intensity and affects the performance of the spectrometer. Therefore, it is necessary to control the film morphology. There are many factors that affect the morphology of thin films, and they are usually divided into external environmental factors and internal characteristic factors. External environmental factors mainly include ambient temperature, substrate temperature, substrate roughness, droplet size, etc., while internal characteristic factors refer to the ink characteristics, including ink surface tension, solute particle size, concentration, viscosity, etc. In terms of ink conditioning, surfactants or adhesives are usually added to regulate the surface tension of the ink solvent. The effect of capillary flow on particles can be overcome by triggering a tension gradient (Marangoni flow) [30] in droplets [31–35], so a uniform and flat film without "coffee rings" can be obtained [36]. Sometimes, increasing the particle size or changing the particle shape can reduce the effect of capillary force and weaken the "coffee ring" effect [37–39]. The solute volume fraction can also be varied to adjust the degree of sparsity or density on deposition patterns [40,41]. As mentioned above, optimizing the film morphology by adjusting the properties of the solvent has excellent effects in printing display panels and preparing visible-light or nearinfrared filter films.

The characteristic spectral peak of HgSe QDs used in this study was located in the longwave infrared at 12.5 µm [42]. The ink cannot be modified simply by adding active agents or polymers when preparing a long-wave infrared filter film. Because most surfactants or adhesives are difficult to evaporate and have strong absorption in the long-wave infrared band, the specific filtering function will not be achieved. Therefore, volatile organic solvents were used to modify the ink in this study. The effects of ink surface tension, particle agglomeration, and the solute volume fraction on the morphology of nanomaterials were investigated.

Eight kinds of evaporable organic solvents (isopropanol, n-octane, ethanol, ethyl acetate, butyl acetate, acetone, chloroform, and toluene) were used as surface tension modifiers. QD inks with different surface tensions were prepared by mixing the organic with toluene–QD solution (toluene was used to ensure the dispersion of QDs). Due to the large difference in polarity between n-octane and toluene, the agglomeration degree of QDs can be regulated by adding n-octane to the QD solution. QD inks with different agglomerated particle sizes could be obtained by mixing different proportions of n-octane and toluene–QD solution. QD inks with different solute volume fractions were prepared by mixing different proportions of toluene and toluene–QD solution. Then, 0.5 µL of the QD solution was dropped on a glass slide with a pipette, simulating the situation of ink droplets on the substrate in inkjet printing. The effects of the surface tension, particle size, and solute volume fraction of the ink solvent on the film morphology were analyzed. The fitting curve of the relationship between the solute volume fraction and transmittance was obtained. The results can provide a reference for the preparation of long-wave infrared QD filter films with specific transmittance and good morphology.

#### **2. Materials and Methods**

The materials and devices used in our experiments are as follows.

The characteristic absorption peak of the QD that we used is 12.5 µm. The QD solution was prepared by dissolving 50 mg of HgSe QDs in 1 mL of toluene. A 50 mg/mL HgSe QD–toluene solution was used as the solute, and isopropanol, n-octane, ethanol, ethyl acetate, butyl acetate, acetone, chloroform, and toluene were used as different solvents. Eight kinds of QD inks with different solvents were obtained by mixing the solute with solvents in a 1:1 volume ratio. Then, the inks were put into an ultrasonic instrument (Kunshan KQ-50B, Beijing, China) and shaken for 10 min. The surface tension of the inks was measured by an automatic surface tensiometer (Zhongchen POWEREACH, Shanghai, China) with the platinum plate method at 11.2 ◦C. Then, 0.5 µL of QD inks were dropped by pipette on glass slides, and the films were observed by using an optical microscope (Mingmei ML31, Guangzhou, China) after the solvent evaporated naturally.

The HgSe QD solution was used as the solute, and the solvent was prepared by mixing toluene and n-octane in volume ratios of 5:5, 4:6, 3:7, 2:8, and 1:9, followed by ultrasonic vibration for 10 min. After that, the inks were obtained by mixing the solvents and the solute with ultrasonic vibration for 10 min. The sampling amount is shown in Table 2.


**Table 2.** Sampling amount.

The surface tension of the inks was measured by an automatic surface tensiometer. Then, 0.5 µL of QD inks was dropped by pipette on glass slides. The films were observed using an optical microscope after the solvent evaporated naturally. The area and number of the particles and the "coffee ring" width were counted and measured using the measurement mode.

The HgSe QD solution was used as the solute, and toluene was used as the solvent. Eight kinds of QD inks with volume fractions *ϕ<sup>µ</sup>* of 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, and 0.75 were obtained by mixing different volumes of toluene with the solute (the effect of QD volume on solute volume was ignored in the calculation), followed by ultrasonic vibration for 10 min, where *ϕ<sup>µ</sup>* = *Va*/(*V<sup>a</sup>* + *V<sup>b</sup>* ), *V<sup>a</sup>* is the volume of QD solution, and *V<sup>b</sup>* is the volume of toluene. Then, 0.5 µL of QD inks was dropped by pipette on glass slides, and the films were observed using an optical microscope and atomic force microscope after the solvent evaporated naturally. Next, 0.5 µL of QD ink was dropped on the ZnSe window using a pipette. The infrared absorption spectrum of the film was measured using a Fourier transform infrared spectrometer (Thermo iS50 FT-IR, Beijing, China) after the solvent evaporated naturally.

All substrates in the experiments were washed three times with acetone, ethanol, and distilled water sequentially and dried in a vacuum desiccator (DZF-6050, Beijing, China).

#### **3. Results**

#### *3.1. Influence of Ink Surface Tension on FILM Morphology*

The morphologies of films prepared with eight different solvent inks as observed under an optical microscope are shown in Figure 1.

**Figure 1.** Morphologies of films prepared with different QD inks. **Figure 1.** Morphologies of films prepared with different QD inks.

It can be seen in Figure 1 that the morphologies of the films prepared with QD inks with different solvents are different, but they can be roughly divided into three types: the first type is multilayer annular deposition (isopropanol and n-octane as solvents); the second type is nonspecific deposition (ethanol, ethyl acetate, and butyl acetate as solvents); and the third type is "gully" deposition (acetone, chloroform, and toluene as solvents). Then, the ink properties were measured and calculated to explore the reasons for film formation. It can be seen in Figure 1 that the morphologies of the films prepared with QD inks with different solvents are different, but they can be roughly divided into three types: the first type is multilayer annular deposition (isopropanol and n-octane as solvents); the second type is nonspecific deposition (ethanol, ethyl acetate, and butyl acetate as solvents); and the third type is "gully" deposition (acetone, chloroform, and toluene as solvents). Then, the ink properties were measured and calculated to explore the reasons for film formation.

The surface tension – of the ink and the droplet radius *R* were measured, and the contact angle of the droplet, the work of adhesion , the work of immersion , and the spreading coefficient were calculated, as shown in Table 3. The surface tension *γg*–*<sup>l</sup>* of the ink and the droplet radius *R* were measured, and the contact angle *θ* of the droplet, the work of adhesion *Wa*, the work of immersion *W<sup>i</sup>* , and the spreading coefficient *S* were calculated, as shown in Table 3.



*S* 0.03 0.05 0.05 0.06 0.07 0.09 0.17 0.32 Due to the small contact angle of the droplet, it cannot be measured by a contact angle meter. However, the droplet can be regarded as a spherical cap with a volume of 0.5 μL, and then the contact angle can be calculated by measuring the droplet radius *R* (Figure Due to the small contact angle of the droplet, it cannot be measured by a contact angle meter. However, the droplet can be regarded as a spherical cap with a volume of 0.5 µL, and then the contact angle can be calculated by measuring the droplet radius *R* (Figure 2a). The calculation formula is

$$V = \frac{\pi h^2}{3}(3r - h) = \frac{\pi h}{6} \left(3R^2 + h^2\right) = 0.5\tag{1}$$

<sup>2</sup> + ℎ 2

The liquid–solid interface wetting of droplets on the substrate can be described by

$$\theta = \arccos \frac{R^2 - h^2}{R^2 + h^2} \tag{2}$$

– = –+– cos (3)

Young's wetting equation [43] (Figure 2b):

3

Therefore, ,

in contact with the substrate [44] (– = −–):

during the contact transition of the three solid–liquid–gas interfaces when the droplet is

= ∆ = – − – − – = −–

= −∆ = – − – − – = –

, and can be calculated from the Gibbs free energy change values

= ∆ = – − – = −– cos (5)

(1 + cos ) (4)

(cos − 1) (6)

**Figure 2.** (**a**) Computational model of spherical cap droplet; (**b**) schematic side view of the contact between the droplet and substrate. **Figure 2.** (**a**) Computational model of spherical cap droplet; (**b**) schematic side view of the contact between the droplet and substrate.

As can be seen in Table 3, the relationship between solvents for the parameters − , , , and was isopropanol < n-octane < ethanol < ethyl acetate < butyl acetate < acetone The liquid–solid interface wetting of droplets on the substrate can be described by Young's wetting equation [43] (Figure 2b):

$$
\gamma\_{s-\mathcal{g}} = \gamma\_{l-s} + \gamma\_{l-\mathcal{g}} \cos \theta \tag{3}
$$

> toluene. The adhesion of droplets to the substrate increased with the increase in | |. The droplets can wet the substrate when ≤ 0, and the wetting ability decreased with the increase in | |. The liquid can spread automatically on the substrate when ≥ 0, and the Therefore, *Wa*, *W<sup>i</sup>* , and *S* can be calculated from the Gibbs free energy change values during the contact transition of the three solid–liquid–gas interfaces when the droplet is in contact with the substrate [44] (*γg*–*<sup>l</sup>* = −*γl*–*<sup>g</sup>* ):

$$\mathcal{W}\_{\mathfrak{d}} = \Delta \mathcal{G} = \gamma\_{\mathfrak{l}-\mathfrak{s}} - \gamma\_{\mathfrak{l}-\mathfrak{g}} - \gamma\_{\mathfrak{s}-\mathfrak{g}} = -\gamma\_{\mathfrak{g}-\mathfrak{l}}(1 + \cos \theta) \tag{4}$$

$$\mathcal{W}\_{\!i} = \Delta \mathcal{G} = \gamma\_{\! -s} - \gamma\_{\! \! -s} = -\gamma\_{\! \! -l} \cos \theta \tag{5}$$

$$\mathcal{S} = -\Delta \mathcal{G} = \gamma\_{\mathcal{S} \to \mathfrak{s}} - \gamma\_{\mathcal{S} \vdash l} - \gamma\_{l \vdash \mathfrak{s}} = \gamma\_{\mathcal{S} \vdash l}(\cos \theta - 1) \tag{6}$$

toluene and chloroform as solvents were higher. Thus, the droplets had stronger adhesion and weaker wetting and spreading abilities on the substrate with no multilayer annular deposition. Therefore, the film-forming property and uniformity can be improved by appropriately increasing the surface tension of the ink. As can be seen in Table 3, the relationship between solvents for the parameters *γg*−*<sup>l</sup>* , *θ*, *Wa*, and *S* was isopropanol < n-octane < ethanol < ethyl acetate < butyl acetate < acetone < chloroform < toluene. The relationship between solvents for the parameters R and *W<sup>i</sup>* was isopropanol > n-octane > ethanol > ethyl acetate > butyl acetate > acetone > chloroform > toluene.

*3.2. Effect of Particle Size on Film Morphology* It was found that the content of the organic solvent in the ink can affect the agglomeration degree of QDs, which in turn affects the morphology of the film. Due to the large difference in polarity, it will cause obvious agglomeration with the addition of n-octane The adhesion of droplets to the substrate increased with the increase in |*Wa*|. The droplets can wet the substrate when *W<sup>i</sup>* ≤ 0, and the wetting ability decreased with the increase in |*W<sup>i</sup>* |. The liquid can spread automatically on the substrate when *S* ≥ 0, and the spreading ability decreased with the increase in *S*.

to the QD–toluene solution. Different agglomerated particles can be obtained by adjusting the ratio between n-octane and toluene. Therefore, inks with volume ratios of toluene to n-octane of 5:5, 4:6, 3:7, 2:8, and 1:9 were prepared. The surface tension, particle size, and Among the above inks, the *θ* of inks with isopropanol and n-octane as solvents was no more than 24.86 mN/m. The droplets had weak adhesion and strong wetting and spreading abilities on the substrate due to the small *Wa*, and the phenomenon of multilayer ring deposition was more likely to occur. In contrast, the *θ* and *W<sup>a</sup>* of the inks with toluene and chloroform as solvents were higher. Thus, the droplets had stronger adhesion and weaker wetting and spreading abilities on the substrate with no multilayer annular deposition. Therefore, the film-forming property and uniformity can be improved by appropriately increasing the surface tension of the ink.

#### *3.2. Effect of Particle Size on Film Morphology*

It was found that the content of the organic solvent in the ink can affect the agglomeration degree of QDs, which in turn affects the morphology of the film. Due to the large difference in polarity, it will cause obvious agglomeration with the addition of n-octane to the QD–toluene solution. Different agglomerated particles can be obtained by adjusting the ratio between n-octane and toluene. Therefore, inks with volume ratios of toluene to

n-octane of 5:5, 4:6, 3:7, 2:8, and 1:9 were prepared. The surface tension, particle size, and "coffee ring" width of the films were measured. The film morphology under the microscope is shown in Figure 3. The particle size distribution in the film is shown in Figure 4. The ink surface tension, film particle size, and "coffee ring" width are shown in Table 4. "coffee ring" width of the films were measured. The film morphology under the microscope is shown in Figure 3. The particle size distribution in the film is shown in Figure 4. The ink surface tension, film particle size, and "coffee ring" width are shown in Table 4. "coffee ring" width of the films were measured. The film morphology under the microscope is shown in Figure 3. The particle size distribution in the film is shown in Figure 4.

The ink surface tension, film particle size, and "coffee ring" width are shown in Table 4.

*Nanomaterials* **2022**, *12*, x FOR PEER REVIEW 6 of 12

*Nanomaterials* **2022**, *12*, x FOR PEER REVIEW 6 of 12

**Figure 3.** The morphologies of films prepared with different QD inks. The volume ratios of toluene to n-octane in the inks are 5:5, 4:6, 3:7, 2:8, and 1:9. **Figure 3.** The morphologies of films prepared with different QD inks. The volume ratios of toluene to n-octane in the inks are 5:5, 4:6, 3:7, 2:8, and 1:9. **Figure 3.** The morphologies of films prepared with different QD inks. The volume ratios of toluene to n-octane in the inks are 5:5, 4:6, 3:7, 2:8, and 1:9.

20

20

25

20

25

**Figure 4.** Particle size distribution diagram. The volume ratios of toluene and n-octane in the inks are 5:5, 4:6, 3:7, 2:8, and 1:9. **Figure 4.** Particle size distribution diagram. The volume ratios of toluene and n-octane in the inks are 5:5, 4:6, 3:7, 2:8, and 1:9. **Figure 4.** Particle size distribution diagram. The volume ratios of toluene and n-octane in the inks are 5:5, 4:6, 3:7, 2:8, and 1:9.

**Volume Ratio of Toluene to N-Octane 5:5 4:6 3:7 2:8 1:9 Volume Ratio of Toluene to N-Octane 5:5 4:6 3:7 2:8 1:9 Table 4.** Ink surface tension, particle size, and "coffee ring" width.

**Table 4.** Ink surface tension, particle size, and "coffee ring" width.

**Table 4.** Ink surface tension, particle size, and "coffee ring" width.


of QDs decreased with the increase in n-octane content. The particle size of the QDs in-

creased, and the "coffee ring" became wider due to agglomeration. It can be seen from the discussion in Section 3.1 that the smaller the surface tension of the ink, the more likely the film has the morphology of multilayer ring deposition. However, the number of "coffee rings" in Figure 3 decreases as the surface tension decreases. This was because the dispersion ability of QDs decreased as the n-octane content increased. It was difficult for the capillary flow in the droplet to push the large particles toward the contact line due to agglomeration. The liquid film evaporated to dryness before the large particles reached the contact line, so the "coffee ring" widened. When the agglomerated particle size of QDs was equal to 188.11 nm, the "coffee ring" and multilayer ring deposition essentially disappeared. When the particle size was equal to 303.89 nm, the large-size particles were primarily concentrated in the center of the film. The "coffee ring" and multilayer annular deposition disappeared completely. Therefore, the film-forming property and uniformity can be improved by appropriately increasing the size of the particle. *3.3. Effect of Solute Volume Fraction on Film Morphology* of QDs decreased with the increase in n-octane content. The particle size of the QDs increased, and the "coffee ring" became wider due to agglomeration. It can be seen from the discussion in Section 3.1 that the smaller the surface tension of the ink, the more likely the film has the morphology of multilayer ring deposition. However, the number of "coffee rings" in Figure 3 decreases as the surface tension decreases. This was because the dispersion ability of QDs decreased as the n-octane content increased. It was difficult for the capillary flow in the droplet to push the large particles toward the contact line due to agglomeration. The liquid film evaporated to dryness before the large particles reached the contact line, so the "coffee ring" widened. When the agglomerated particle size of QDs was equal to 188.11 nm, the "coffee ring" and multilayer ring deposition essentially disappeared. When the particle size was equal to 303.89 nm, the large-size particles were primarily concentrated in the center of the film. The "coffee ring" and multilayer annular deposition disappeared completely. Therefore, the film-forming property and uniformity can be improved by appropriately increasing the size of the particle. *3.3. Effect of Solute Volume Fraction on Film Morphology* It can be seen in Table 4 that the surface tension of the ink and the dispersion ability of QDs decreased with the increase in n-octane content. The particle size of the QDs increased, and the "coffee ring" became wider due to agglomeration. It can be seen from the discussion in Section 3.1 that the smaller the surface tension of the ink, the more likely the film has the morphology of multilayer ring deposition. However, the number of "coffee rings" in Figure 3 decreases as the surface tension decreases. This was because the dispersion ability of QDs decreased as the n-octane content increased. It was difficult for the capillary flow in the droplet to push the large particles toward the contact line due to agglomeration. The liquid film evaporated to dryness before the large particles reached the contact line, so the "coffee ring" widened. When the agglomerated particle size of QDs was equal to 188.11 nm, the "coffee ring" and multilayer ring deposition essentially disappeared. When the particle size was equal to 303.89 nm, the large-size particles were primarily concentrated in the center of the film. The "coffee ring" and multilayer annular deposition disappeared completely. Therefore, the film-forming property and uniformity can be improved by appropriately increasing the size of the particle.

#### It was found that toluene had the best dispersing effect on QDs. Therefore, toluene It was found that toluene had the best dispersing effect on QDs. Therefore, toluene *3.3. Effect of Solute Volume Fraction on Film Morphology*

was used as the solvent, and QD inks with solute volume fractions of 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, and 0.75 were prepared. The thin films were obtained by evaporation. The morphologies of QD films were observed with an optical microscope, as shown in Figure 5a. was used as the solvent, and QD inks with solute volume fractions of 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, and 0.75 were prepared. The thin films were obtained by evaporation. The morphologies of QD films were observed with an optical microscope, as shown in Figure 5a. It was found that toluene had the best dispersing effect on QDs. Therefore, toluene was used as the solvent, and QD inks with solute volume fractions of 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, and 0.75 were prepared. The thin films were obtained by evaporation. The morphologies of QD films were observed with an optical microscope, as shown in Figure 5a.

**Figure 5.** (**a**) The morphologies of films prepared by inks with solute volume fractions of 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, and 0.75 under the microscope. (**b**) The evaporation process when the solute volume fraction was 0.5. (**c**) The schematic side and top views of the liquid film evaporation process when the solute volume fraction was 0.5. **Figure 5.** (**a**) The morphologies of films prepared by inks with solute volume fractions of 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, and 0.75 under the microscope. (**b**) The evaporation process when the solute volume fraction was 0.5. (**c**) The schematic side and top views of the liquid film evaporation process when the solute volume fraction was 0.5.

The experimental results show that the solute volume fraction was not the factor that determined the formation of the "coffee ring". This was only determined by the characteristic of the solvent. Since the solute cannot unpin the contact line and redirect the flow, the evaporation rate of the edge of the liquid film was greater than that of the center when the solvent was constant. In order to keep the contact line pinned, there must be a continuous, radially outward capillary flow from the center to the contact line to compensate for

When the solute volume fraction was less than 0.1, some QDs were deposited on the glass slide before they moved to the edge due to the flash evaporation rate of the solvent and finally formed a uniform QD film. When the solute volume fraction was greater than 0.1 (for example, *φ<sup>μ</sup>* of 0.5), the film evaporation process was more complicated and formed a gully-like morphology, as shown in Figure 5b. The schematic side and top views

the evaporative removal of the liquid, eventually forming a "coffee ring".

of the liquid film evaporation process are shown in Figure 5c.

As can be seen in Figure 5b, when the ink contacts the substrate, it forms a liquid film. Then, the three-phase contact line is pinned at once. The QDs in the liquid film began to move to the three-phase contact line. At this time, the evaporation mode was the constant contact radius model (CCR). When the evaporation proceeded for 15 s, the reverse The experimental results show that the solute volume fraction was not the factor that determined the formation of the "coffee ring". This was only determined by the characteristic of the solvent. Since the solute cannot unpin the contact line and redirect the flow, the evaporation rate of the edge of the liquid film was greater than that of the center when the solvent was constant. In order to keep the contact line pinned, there must be a continuous, radially outward capillary flow from the center to the contact line to compensate for the evaporative removal of the liquid, eventually forming a "coffee ring".

When the solute volume fraction was less than 0.1, some QDs were deposited on the glass slide before they moved to the edge due to the flash evaporation rate of the solvent and finally formed a uniform QD film. When the solute volume fraction was greater than 0.1 (for example, *ϕ<sup>µ</sup>* of 0.5), the film evaporation process was more complicated and formed a gully-like morphology, as shown in Figure 5b. The schematic side and top views of the liquid film evaporation process are shown in Figure 5c.

As can be seen in Figure 5b, when the ink contacts the substrate, it forms a liquid film. Then, the three-phase contact line is pinned at once. The QDs in the liquid film began to move to the three-phase contact line. At this time, the evaporation mode was the constant contact radius model (CCR). When the evaporation proceeded for 15 s, the reverse "coffee ring" 'a' appeared, gradually widened, and moved toward the center of the liquid film. At

the same time, some QDs in ring 'a' diffused toward the "coffee ring" under the action of capillary force. The "coffee ring" continued to widen. When the evaporation proceeded for 40 s, the liquid film was released from the pinned "coffee ring", and the short-term constant contact angle (CCA model) evaporation mode occurred. The pinned ring 'b' was the new three-phase contact line, and the film continued to evaporate in CCR mode. As the liquid film gradually became thinner, the temperature difference between the edge and center of the liquid film became smaller, and the moving speed of ring 'a' to the center slowed down. When the evaporation progressed to around 70% (at 50 s), ring 'a' was fixed and flushed out within 5 s. At the same time, the stably distributed QDs in the center of the liquid film also began to move rapidly to the edge. The liquid film was too thin to be fixed after 15 s, so it shrunk rapidly toward the center and evaporated to dryness. QDs piled up on the edges, eventually forming a "gully" morphology of varying depths. Therefore, in order to avoid the appearance of the "gully" and obtain a more uniform and flat QD film, the volume fraction of the ink solute should not be greater than 0.1.

#### *3.4. Infrared Transmittance Analysis of Thin Films*

In order to analyze the relationship between film morphology and film transmittance and to provide an empirical method for the subsequent preparation of long-wave infrared QD films, the film morphology was characterized by atomic force microscopy. The 3D surface topography was recorded using a Nanosurf Flex-Axiom atomic force microscope (Nanosurf, AG) in soft tapping mode with a scan speed of 6.25 µm/s to obtain 104 × 104-pixel images. The experiments were carried out at room temperature (297 ± 1 K) using cantilevers with the following nominal properties for force–distance curve measurements: a length of 125 µm, a width of 25 µm, a thickness of 2.1 µm, a tip radius of 10 nm, a force constant of 5 N/m, and a resonance frequency of 150 kHz, as shown in Figure 6.

It can be seen in Figure 6a,b that the QDs are distributed in islands on the substrate. The QDs became denser and higher with the increase in the solute volume fraction. It can be seen in Figure 6c that the shape of the QDs appears broader, and the cross-sectional diameter became larger with the further increase in the solute volume fraction.

The arithmetic mean heights (Sa) of films with solute volume fractions of 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, and 0.75 were 53.90 nm, 55.25 nm, 59.23 nm, 61.83 nm, 66.13 nm, 66.82 nm, ands 72.25 nm, respectively, as shown in Figure 7a. The transmissions of films with solute volume fractions of 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, and 0.75 were 88.65%, 81.27%, 61.91%, 59.14%, 45.36%, 37.38%, and 33.85%, respectively, as shown in Figure 7b. The fitting curve of the film height and transmission is shown in Figure 7c.

It can be seen in Figure 7c that the increase rate of the height of the film and the decrease rate of the transmission at the characteristic peak became slower when the solute volume fraction was 0.1. There was a linear relationship between the height and transmission, which conformed to the Lambert–Beer law. This result can provide an important reference for the preparation of thin films with specific transmission.

Transmission (%)

Heigh(nm)

*3.4. Infrared Transmittance Analysis of Thin Films*

**Figure 6.** (**a**) Contour map of the film. (**b**) Three-dimensional images of the film under AFM. (**c**) Sectional view of the film's diagonal. (The volume fractions of ink solute are 0.025, 0.075, 0.25, and 0.75.). nm, ands 72.25 nm, respectively, as shown in Figure 7a. The transmissions of films with solute volume fractions of 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, and 0.75 were 88.65%, 81.27%, 61.91%, 59.14%, 45.36%, 37.38%, and 33.85%, respectively, as shown in Figure 7b. The fitting curve of the film height and transmission is shown in Figure 7c.

"coffee ring" 'a' appeared, gradually widened, and moved toward the center of the liquid film. At the same time, some QDs in ring 'a' diffused toward the "coffee ring" under the action of capillary force. The "coffee ring" continued to widen. When the evaporation proceeded for 40 s, the liquid film was released from the pinned "coffee ring", and the shortterm constant contact angle (CCA model) evaporation mode occurred. The pinned ring 'b' was the new three-phase contact line, and the film continued to evaporate in CCR mode. As the liquid film gradually became thinner, the temperature difference between the edge and center of the liquid film became smaller, and the moving speed of ring 'a' to the center slowed down. When the evaporation progressed to around 70% (at 50 s), ring 'a' was fixed and flushed out within 5 s. At the same time, the stably distributed QDs in the center of the liquid film also began to move rapidly to the edge. The liquid film was too thin to be fixed after 15 s, so it shrunk rapidly toward the center and evaporated to dryness. QDs piled up on the edges, eventually forming a "gully" morphology of varying depths. Therefore, in order to avoid the appearance of the "gully" and obtain a more uniform and flat QD film, the volume fraction of the ink solute should not be greater than 0.1.

In order to analyze the relationship between film morphology and film transmittance and to provide an empirical method for the subsequent preparation of long-wave infrared QD films, the film morphology was characterized by atomic force microscopy. The 3D surface topography was recorded using a Nanosurf Flex-Axiom atomic force microscope (Nanosurf, AG) in soft tapping mode with a scan speed of 6.25 μm/s to obtain 104 × 104 pixel images. The experiments were carried out at room temperature (297 ± 1 K) using cantilevers with the following nominal properties for force–distance curve measurements: a length of 125 µm, a width of 25 µm, a thickness of 2.1 µm, a tip radius of 10 nm, a force

constant of 5 N/m, and a resonance frequency of 150 kHz, as shown in Figure 6.

**Figure 7.** (**a**) Transmittance of films prepared from QD inks with different solute volume fractions. (**b**) Height of films prepared from QD inks with different solute volume fractions. (**c**) The fitting function of solute volume fraction and film transmittance. **Figure 7.** (**a**) Transmittance of films prepared from QD inks with different solute volume fractions. (**b**) Height of films prepared from QD inks with different solute volume fractions. (**c**) The fitting function of solute volume fraction and film transmittance.

It can be seen in Figure 7c that the increase rate of the height of the film and the decrease rate of the transmission at the characteristic peak became slower when the solute

HgSe QD inks with characteristic spectral bands located in the long-wave infrared were prepared by mixing various organic solvents with a toluene solution of QDs. Among them, QD inks with different tensions were first obtained by mixing eight kinds of organic solvents (isopropanol, n-octane, ethanol, ethyl acetate, butyl acetate, acetone, chloroform,

**4. Conclusions**

reference for the preparation of thin films with specific transmission.

#### **4. Conclusions**

HgSe QD inks with characteristic spectral bands located in the long-wave infrared were prepared by mixing various organic solvents with a toluene solution of QDs. Among them, QD inks with different tensions were first obtained by mixing eight kinds of organic solvents (isopropanol, n-octane, ethanol, ethyl acetate, butyl acetate, acetone, chloroform, and toluene) with the QD–toluene solution. Secondly, due to the large difference in polarity, QD inks with different agglomerated particle sizes were obtained by mixing different proportions of n-octane and QD–toluene solution. QD inks with different solute volume fractions were then prepared by mixing different proportions of toluene and toluene–QD solution. Finally, films with different morphologies were obtained by naturally evaporating QD ink droplets on the substrate. The effects of the surface tension, particle size, and volume fraction on the film morphology were emphasized in the analysis. After that, the infrared transmission spectra of the films were measured. The experimental results suggest that the film slipped in the evaporation process, and the multilayer annular deposition formed when the surface tension of the ink was no more than 24.86 mN/m. The "coffee ring" and the multilayer annular deposition essentially disappeared when the solute particles were larger than 188.11 nm. When the solute volume fraction was greater than 0.1, the QDs in the film were accumulated, and a "gully" morphology appeared. In addition, the increase rate of the film height and the decrease rate of transmission slowed down. The relationship between the film height and transmission was fitted, and the curve conformed to the Lambert–Beer law. Therefore, the morphology of the film can be improved by adjusting the surface tension of the film, the particle size of the solute, and the volume fraction of the solute. Therefore, a uniform and flat film without "coffee rings" can be prepared by adjusting the surface tension, particle size, and volume fraction. This approach could provide an empirical method for the preparation of LWIR QD filter film arrays. It was also found that the evaporation rate, temperature or type of substrate, and shape of solute particles also affected the film morphology in the experiment. The above factors can be discussed and analyzed in detail in subsequent research. In addition, agglomeration easily occurs due to the large specific surface area of nanoparticles. Therefore, it is also important to modify and passivate the surface to avoid agglomeration when synthesizing nanomaterials, and QD filter films with good morphology can be prepared by improving the ink uniformity.

**Author Contributions:** S.W.: methodology, validation, investigation, and writing—original draft; X.Z.: methodology; Y.W.: AFM test and writing—review; T.G.: writing—review and editing; S.C.: supervision and writing—review and editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work is funded by the National Key R&D Program of China (2021YFC3330201).

**Data Availability Statement:** Not applicable.

**Acknowledgments:** Thanks to Yi Wang for technical support in AFM testing. Thanks to Yong Pan for the theoretical support for calculations.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

